NASA TN D-206
M^iA
TEGHNIGAL NOTE
D-206
HETEROGENEOUS COMBUSTION OF MULTICOMPONENT FUELS
By Bernard J. Wood, Henry Wise, and S. Henry Inami
Stanford Research Institute
NATIONAL AERONAUTICS AND SPACE ADMINISTRATION
WASHINGTON November 1959
(liASA-TH-D-206) f.ElEfiOGBMEClJ£ COBBOSIIOB Of
tCLiicciifCNEKT iDELS (Stanlord Research
Ir.£t.) 36 p
00/25
M83-710C2
ODclas
0197738
NATIONAL AERONAUTICS AND SPACE AIMINISTRATION
TECHNICAL NOTE D-206
HETEROGENEOUS COMBUSTION OF MULTICOMPONENT FUELS
?y Bernard J. Wood, Henry Wise, and S, Henry Inami
SUMMARY
The use of multicomponent fuels in modern propulsion systems led
to this study of the combustion characteristics of a fuel drop containing
liquid constituents of different physical and chemical properties. Direct
measurements of the rate of change of cross-sectional area of a burning
drop were obtained with a photoelectric shadowgraph apparatus . The varia-
tion of burning-rate coefficient as a function of drop composition, was
determined for a group of binary fuel mixtures . In addition, the burning
characteristics of drops of hydrocarbon fuels containing up to five con-
stituents and of industrial fuel mixtures, such as kerosene and jet fuel
JP-14- , were studied .
The experimental results indicate that during combustion the composi-
tion of a multicomponent fuel drop changes by a process of simple batch
distillation; that is, the vapors produced are removed continuously with-
out further contact with the residual liquid mlxtiure. Hence, as burning
proceeds the more volatile constituents of the liquid drop vaporize first
and the concentration of the higher boiling fractions in the liquid phase
increases . Temperature and composition gradients within the burning drop
appear to be modified by internal circulation of the fuel.
For a given fuel mixture, the stability of the drop during the com-
bustion process was found to be a function of the initial drop size, the
nature and relative quantities of the components in the liquid phase,
and the magnitude of the differences in their boiling points . Motion
pictures of drops of mixt\rres which burn disruptively revealed the forma-
tion of bubbles which expand and burst in the liquid phase. It is sug-
gested that these effects are due to radiative heat transfer from the
flame envelope to the liquid, which results in local vaporization of the
more volatile components within the drop . It can be shown that the
radiative heat flux from the flame to the liquid is independent of drop
size; the conductive heat flux, however, is inversely proportional to
drop diameter. Thus, with decreasing drop size the latter mechanism of
heat transfer begins tc make an increasing contribution. If the rates
of diffusion of mass and energy within the liquid are slow relative to
the mass burning rate, nonequilibrium distillation is to be expected.
W-129
The surface temperatures of burning drops of pure compounds were
measured for (l) A drop suspended from a thermocouple Jianction^ and
(2) a fuel-wetted, porous, Alundum sphere. The values obtained from
these measiu-ements are in close agreement with those calculated on the
basis of a spherlcosymmetric model of a burning drop.
INTROIXJCTION
During the past few years a large amo\mt of information on the
heterogeneous ccmbustion of drops of pure fuels has been obtained
(ref. l) . Liquid fuels of industrial and military importance, however, ^
are generally composed of mixtures of chemical ccHirpounds possessing ^
individually different physical and chemical properties. The burning
characteristics of such mult i component mixtures when dispersed as liquid
aerosols have not been investigated in detail. Combustion of a binary
fuel mixture over its free liquid siorface produces a gradual cheuige in
the composition of the liquid phase (ref. 2); on the other hajid, com-
bustion of a fuel mixtiore on a wick immersed in the fluid consumes the
fuel without changing its liquid composition (ref. 5)- No deductions
concerning the behavior of a mult i component burning drop can be made
frcxa these observations, since in such a system the conductive heat
flxix increases with decreasing drop size, and the convective effects
in the liquid phase are affected by the geometry.
Because of the small mass of fluid involved, it is difficult to
measiore the change in the chemical composition of a burning, multi-
component fuel drop. Therefore, the variation in the burning-rate coef-
ficient e was chosen as a means of examining the type of vaporization
process occurring on the drop surface. This coefficient represents the
slope of the ciirve depicting the time rate of change in cross- sectional
area of the drop during combustion. For pure fuels the bxirning-rate
coefficient remains constant (ciorves a and b, fig. l) during the life-
time of the drop (ref. l) . Similarly, for a binaxy fuel mixture com-
posed of compounds with individiially different burning-rate coefficients,
the drop cross-sectional area decreases linearly with time if the ccan-
position of the liquid phase does not vary during combustion (curve c,
fig. l) . A fuel drop, however, whose constituents exhibit different
burning-rate coefficients (curves e and f, fig. l) and vaporize selec-
tively in the manner of a batch distillation dioring ccmbustlon, decreases
in cross-sectional area nonlinearly with respect to time (curve g, fig. l).
Consequently, measurement of the burning-rate coefficients of binary
mixtures offers a suitable means of analyzing the vaporization process ^
occurring during combustion.
The present investigation was conducted at Stanford Research
Institute under the sponsorship and with the financial assistance of
the National Advisory Committee for Aeronautics.
SYMBOLS
c specific heat at constant pressure of liquid at mean tem-
perature between Tg^ and Tg, cal/g °C
Cg specific heat at constant pressure of air at temperature Tg,
cal/g °C
D drop diameter, cm
i stoichiometric mixture ratio, {y&l^f) f -t y,
L heat of vaporization (sensible heat plus latent heat), cal/g
m mass burning rate of fuel, g/sec
n total niamber of moles of fuel comprising drop at t^
n' total number of moles of fuel comprising drop at t
N mole fraction of individual coiirponent in liquid phase ot
burning drop
N' mole fraction of individual component in liquid phase of
bijrning drop at t
N^ mole fraction of individual component in vapor phase in
equilibrium at interface of burning drop
q heat of combustion (fuel vapor to vapor products), cal/g
R drop radius, cm
t time, sec
tQ instant of ignition, sec
tf. instant at which fuel drop is totally consigned, sec
Tg^ temperature of ambient air, °K
Tc adiabatic flame temperature, °K
c "" a o
T„ log mean temperature, ; ; — -, K
2.5 log {TjT^y
Tg temperature at drop siurface, K
Yg^ weight fraction of oxygen in air, 0.232
Yf. weight fraction of fuel in vapor at interface of biirning drop
e b\irning-rate coefficient, (diameter) /sec, cm^/sec
A thermal conductivity of air at temperature Tg, cal/sec cm °C
p density of liquid, g/caP
APPARATUS
For liquid fuel drops sufficiently small that the effect of con-
vective heat transfer can be neglected, the mass burning rate is pro-
portional to the drop diameter (refs. k and 5). Consequently, the sior-
face area of the burning drop decreases linearly with time.
d2 = et.
t - (t)]
(1)
where the b\irning-rate coefficient e is given by
€ = itHL (2)
rtpD
In this investigation, burning-rate coefficients were determined
by means of a photoelectric shadowgraph apparatus which directly meas-
ures the variation of cross-sectional area of a single drop as a func-
tion of time. The device was calibrated to give the cross-sectional
area of a drop in terms of the square of the diameter of an equivalent
sphere. The essential features of the apparatus are illustrated in
figure 2 and are described in detail In reference 7. Fuel drops
employed in the measurements had an initial diameter of the order of
1,800 microns and were burned in air at a pressure of 1 atmosphere.
They were suspended from a quartz sphere approximately 8OO microns in
diameter which was formed on the end of a quartz fiber with a diameter
of 80 microns . The drops were ignited with the flame of an alcohol
burner .
Siorface temperature measiorements of pure fuel drops were carried
out by means of a fuel-wetted porous sphere, which has previously been
W described (ref. 5)- The device was modified as shown in figure 3- A
1 surface temperature measurement was made by adjusting the fuel-feed rate
2 to the sphere to exceed the mass burning rate of the system, thus estab-
9 lishing a surplus of liquid at the sphere sxjrface. This surplus coa-
lesced in a pendent mass of fluid which completely enveloped the thermo-
couple junction (Chromel-Alumel, 3-niil wires). When the fuel- feed rate
was decreased to a value less than the characteristic mass burning rate,
the mass of pendent fluid diminished, and its receding lower surface
advanced toward the junction. The moment the thermocouple junction
pierced the film of liquid it crossed a large temperature gradient, and
the slope of the thermocouple output curve traced by a recording poten-
tiometer increased sharply. The temperature indicated at this departure
was taken as the surface temperatiore of the burning drop.
Additional surface temperature measurements were made by biirning
small drops of fuels suspended from the junction of a thermocouple. The
thermocouple wires were insulated from the hot zone of the flame by a
noncatalytic ceramic coating, National Bureau of Standards ceramic
coating A-ii-lB obtained from California Metal Enameling Co. The thermo-
couple junction was situated about 1 millimeter below a ceramic bead
fused to the wires, which served as a supporting siorface for the liquid
drop. A drop suspended from this fixtiire assumes a lemon shape, for
the pressiore of the thermocouple junction against the surface causes a
slight protuberance . The liquid- vapor interface remains integral, how-
ever, during the burning lifetime of the drop; the thermocouple jixnction
pierces the surface only at the moment the drop is totally consumed.
The temperature, displayed as a function of time on an oscilloscope,
gradually rises to a steady value which is terminated by a sudden, sharp
increase (fig. h) . The initial rise obviously represents the warming-up
of the drop and possibly its support, while the final spurt marks the
penetration of the drop surface by the thermocouple junction; the tem-
perature indicated by the level portion of the curve was selected as
the value of the surface temperature .
EXPERIMENTAL RESULTS
The three binary fuel mlxtiires studied In this Investigation were:
(l) Heptane and butanol-1, (2) 2,2,i<-trimethylpentane and butanoI-1, and
(5) dibutyl ether and pentanol-1. The alcohols were analytical reagents
used without further purification; the hydrocarbons were obtained from
Phillips Petroleum Co., and were rated at 99'^-percent purity; the dibutyl
ether was Eastman practical grade. Mlxtiires of various proportions of
these liquids were prepared by volumetric dispensation of the components
from burettes . W
1
In figure 5 the drop burning characteristics of binary mixtures are 2
compared with those of their pure constituents . The pure fuel drops 9
exhibit linear changes of cross-sectional area with time, that is, their ^
burning-rate coefficients remain constant during the drop lifetime. On
the other hand, the biirning-rate curve produced by the fuel mixture con-
taining equimolar quantities of butanol and heptane, for example, exhib-
its curvature. At the beginning of combustion the slope approaches
that of the azeotroplc composition (25 mol-percent butanol), while near
the end, that of pure butanol. The initial and final values of the
b\irning-rate coefficients of the various fuel mixtures are summarized
in figure 6. The results Indicate that the composition of the liquid
phase changes during combustion in a manner which can be described as
a batch distillation process, in which the liquid and the vapor are in
equilibrium with respect to composition ajid temperatiore, rather than
nonequilibrium distillation (see calculated values indicated in figure 6).
For mixtures which contain a relatively small mole fraction of their more
volatile components, the experimental data suggest constant bxirning rates.
This effect is probably due in part to the loss of the volatile component
from the liquid phase preceding ignition of the drop and to the ignition
process .
The drop burning characteristics of mixtures of a number of ali-
phatic hydrocarbons with carbon chains varying between seven and sixteen
atoms in length were observed. A number of these mixtures exhibited an
\msteadlness during combustion which is termed disruptive combustion
(table I). Visually, the phenomenon appears as a sudden oscillation of
the drop on its supporting fiber, accompanied by ejection of small q\ian-
tities of liquid fuel in random directions . . The manner in which disrup-
tive combustion affects the burning ciorve is demonstrated in the oscil-
loscope record in figure 7-
The measured surface temperatures of burning drops are summarized
in table II. There is good agreement between the data obtained by the
two experimental methods . ••
DISCUSSION
Combustion of a Miiltl component Fuel Drop
The burning rate of a small liquid fuel drop in an oxidizing
atmosphere is controlled by the transport of mass and energy. Conse-
quently it is a function of the thermodynamic properties of the fuel-
oxidizer system and the transport properties of the gaseous envelope
surroimding the drop. Since these parameters are characteristics of
the chemical composition of a fuel, it is evident that the burning rate
of any mixture of mlscible compounds depends upon the nature and pro-
portions of the constituents.
There are two basic mechanisms by which fuel can vaporize from the
surface of a burning drop made up of a multicomponent solution. First,
the vapors leaving the surface may have a composition identical to that
of the liquid phase, as shown schematically in figure 8, line A. Under
such conditions of nonequilibrium distillation, the burning character-
istics of a binary fuel mixture resemble those of a pure fuel, for there
occurs no change of composition during the lifetime of the drop.
Secondly, a thermodynamic equilibrium may be established at the
interface between liquid and vapor so that the components vaporize from
the burning drop in continuously varying composition ratios, each of
which is loniquely determined by the instantaneous temperature of the
liquid phase (fig. 8, line B) , It is conceivable also that the vapor-
ization process may occur in a manner intermediate between these two
cases, producing an unpredictable variation in the composition of the
\inburned portion of the drop (fig. 8, line C) . These processes assume
a \iniform composition within the liquid at every instant, such as coiold
be produced by convective mixing within the drop. It has been found by
El Wakll and co-workers (ref . 8) that the fluid within a drop evaporating
in a heated airstream circulates rapidly enough to eliminate temperature
gradients in the liquid. Similar observations, which demonstrated cir-
culation within the liquid of a burning drop, were made during this
investigation.
For a burning-drop model in which a spherical flame envelope is
maintained around the liquid sphere, heat- and mass- transfer considera-
tions lead to an expression of the Isurning-rate coefficient in terms of
thermodynamic properties of the drop and transport properties of the gas
film between drop and flame . Since the gaseous region between the fuel
and the reaction zone of a diffusion flame has been shown to consist pre-
dominantly of nitrogen, carbon dioxide, and water vapor for a number of
representative organic fuels (ref. 9, PP' 759 and 1059), the transport
properties of the gaseous envelope around a burning drop may be con-
sidered independent of liquid composition within the category of ordinary
8
organic fuels. In addition, it was noted that the quotient of heat of
combustion and stoichiometric mixture ratio (q/i) for the fuels employed
in this investigation is essentially a constant (table III) . At any
instant during the combustion process the theoretical burning-rate coef-
ficient for the binary mixtiire is:
8A,
%2 - (YiPi + Y2P2)c,
log.
1 +
(Yl^l
Y2L2)
[¥
>(Ts - Ta)
(3)
Burning-rate coefficients for drops of the pure constituents of the
mlxtvures were calculated from this equation, using the numerical values
given in table III. The adiabatic flame temperatures of these fuels are
nearly identical; so for all cases Ag and Cg were evaliiated for air
at the logarithmic mean temperature between T^ and Tg^ (ref. 6).
An examination of the burning curves in fig\ires 5(a) and 5(^) shows
that fuel mixtures in the center of the composition range display time
variations in their burning rates. A theoretical analysis was attempted,
based on the assiomption that the composition of the drop at any time
during the burning lifetime could be characterized from the equilibrium
distillation curves for the appropriate system. Such an assumption
requires that (l) The ignition lag of the drop be small, so that the
liquid-vapor interface rapidly attains thermodynamic equilibrium with
its immediate surroundings, and (2) the temperature of the interface be
sufficiently close to the normal boiling point of the system (cf . table II)
to permit composition to be estimated from ordinary distillation curves
for 1 atmosphere pressure.
The burning curves obtained from drops of pure organic compounds
and binary azeotropes give an indication of the ignition lag. The
instant of ignition of the drops is indicated in figure 5 by the arrows,
with an estimated precision of ±0.1 second. The shoulder of each curve
immediately follows ignition and spans the period of time during which
the burning rate of the drop rises to its characteristic steady-state
value. In all cases, the length of this span represents approximately
15 percent of the total biirning lifetime of the drop.
In accord with theoretical calculations, experimental measixrements
on burning drops of the pure compounds used in this investigation demon-
strated that the surface temperatures are below the normal boiling
points of the fuels (table II) . Consequently, the estimation of the
time variation of composition of a drop on the basis of ordinary dis-
tillation ciirves can be regarded only as a first approximation.
The distillation curves for the three systems of interest were
determined in this laboratory. Since liquid- vapor equilibrium data for
these binary mixtures are not presently available in the literature, the
phase diagrams and the techniques employed in their determination are
presented in the appendix.
The Rayleigh equation
log^
(Nv - N)
W relates the composition of a residue in a simple batch distillation to
1 the mole fraction of the original dlstilland remaining at any given
2 instant (ref . 10) . The value of the Integral between various axbi-
9 trarily selected limits of liquid composition (N to N') was determined
, graphically by summing the area under the plot of N versus i/(Nv - n),
the points for which were obtained from the distillation curves. The
results of this integration were a series of composition values cor-
responding to mole fractions of the liquid drop yet unburned. The
unburned mole fraction of fuel is a function of the cross-sectional
area of the drop. By assuming an average molecular weight for the fuel
mixture, the composition of the liquid phase as a function of time
after ignition may be explicitly calculated from the experimental
burning-rate c\arves. Sets of composition values were obtained, there-
fore, corresponding to periods of time elapsed after ignition for spe-
cific b\arning drops of binary fuels. Examples of these results, which
illustrate the rapidity with which certain mixtures change composition,
are presented in table IV. It is to be noted that the mixture with an
Ititial composition of 10 percent butanol approaches pure 2,2,4-
trimethylpentane as it burns, while the other mixtures approach pure
butanol. This is caused by the existence of an azeotrope with a mini-
mum boiling point at a composition of 17-5 percent butanol.
By introducing the composition values thus obtained into equa-
tion (5), the instantaneous burning-rate coefficients for biirning drops
of binary mixtures of interest were calculated. These values are pre-
sented in the form of theoretical burning-rate curves in fig\ire 9.
Superimposed on the theoretical curves are points taken from the burning-
rate measurements of drops of identical initial composition. In view of
the approximations made in selecting the vapor film properties, the
degree of agreement is satisfactory. One may conclude that after a
brief Ignition lag a simple equilibrium batch distillation is an ade-
quate description of the vaporization process which a two-component
drop undergoes during combustion.
In the light of these experiments it is not surprising that the
burning curves for drops of many multlcomponent fuels of practical
importance are quite linear. The explanation lies in the fact that the
thermodynamic characteristics of all the components are such that the
10
individual burning-rate coefficients vary over a narrow range. The
pentanol-dibutyl ether system, figures 5(c) and 9(c), displays this
property for the case of a binary fuel mixture; the practical fuels,
kerosene and jet fuel JT-k (fig. 10), exhibit this characteristic for
mi Iti component mixtures.
Disruptive Drop Combustion
By photographing burning drops with a motion-picture camera, the
phenomenon of disruptive combustion was shown to be associated with the W
formation of minute bubbles or vapor pockets in the liquid phase . The 1
photographs of figure 11 are successive frames from a motion-picture 2
film which reveal the formation and disintegration of such a disturb- ^
ance. Frame I shows the quietly burning drop 1.1^4-9 seconds after igni-
tion; the dark spots are highlights produced by the strong side illumina-
tion of the drop. The beginning of a bubble is visible in the upper
right quadrant of the drop in frame II, and the subsequent growth of
the disturbance can be followed in frames III through V. The bubble
apparently burst before frame VI was exposed, and within the time
elapsed during the next two frames, the drop returned to quiescent com-
bustion. The burning curve (fig. 7) for "the same drop photographed in
this sequence clearly shows the disruptions. In this case, a small
loss of material from the liquid phase had a negligible effect on the
characteristic burning rate.
It is probable that this phenomenon is associated with the transfer
of radiant energy from the flame into the interior of the drop. Calcu-
lations of temperature profiles have indicated that under conditions of
absorption of incident radiant energy the temperat\ire inside a drop can
exceed that on the surface (ref. 11). The contribution of radiative
heat transfer to a burning drop was demonstrated experimentally by
enhancing the absorptivity of the fuel. Unfortunately, compounds such
as acetonitrile and phenylacetonitrile, which absorb in the spectral
region in which hydrocarbon flames emit strongly (~J+.5 microns), are
quite insoluble in fuels such as those employed in this investigation.
The addition of a particulate black dye which forms a slvurry appeared
to be the only way to change the absorptivity of the hydrocarbon fuel
drops significantly (ref. 12) . To magnify the effect which a change in
fuel absorptivity would produce in a b\arning drop, a fuel mixture was
selected which burned quietly but which possessed a wide boiling range
and therefore a susceptibility to disruptive combustion (cf. table I).
Small quantities of Apiezon W black wax were added to a 50-50-'V"olume-
percent mixture of nonane and tetradecane. Microscopic examination of
the filtered solution, which was brown in color, indicated that the
coloring matter existed in the form of a particulate suspension within
the drop. Absorption spectrograms of the colored solution showed a
very small, uniform increase in optical density over the clear fuel in
11
the region of ^ to 5 microns. Nevertheless, drops of the colored solu-
tion burned disruptively . To preclude the possible explanation that
the particulate material within the drop provided nuclei for bubble
formation, a parallel experiment was caxried out in which finely ground
alumina was dispersed in the fuel. The addition of these white reflec-
tive particles had no effect on quiescent combustion of the nonane-
tetradecane mixture .
It appears, therefore, that radiant energy transfer to a liquid
drop may result in the formation of a local hot zone within the liquid
W phase. If the drop is composed of fuels which boil over a sufficiently
1 large temperature range and which are blended in a suitable quantitative
2 ratio, this hot zone will cause the lower-boiling constituents to vapor-
9 ize rapidly and disrupt the liquid- vapor Interface. Apparently, internal
circulation is not sufficient to eliminate these inhomogeneities caused
by radiant heat transfer. Also, it is to be noted that, although radia-
tive heat flux is independent of drop size, conductive heat flux from
the flame to the liquid increases with decreasing drop diameter. Thus,
the process of batch distillation observed in the present measurements
may change to one of nonequilibrium vaporization for drops much smaller
than those employed in these studies. Under these conditions, the rates
of convective transfer of mass and. energy within the liquid phase may
be slow relative to the mass-burning rate. The diminished role of
radiative heat transfer in small drops would decrease the likelihood
of disruptive burning; but the diminution of convective heat and mass
transfer within the liquid phase would augment the tendency of a drop
to burn disruptively as a result of increased conductive heat flux.
Hence the dependence of disruptive combustion on drop diameter cannot
be predicted quantitatively .
An estimate of the disrupt ibility of fuel drops in a combustion
chamber should not be based entirely on the burning characteristics of
single drops, for, in practice, a major portion of the radiant heat
seen by a drop arises from the hot walls of the chamber. The spectral
characteristics of such hot-wall radiation are quite different from the
energy spectrum of a flame.
Surface Temperature of Drop
Calculations of the drop s\jrface temperature are based on the
burning-rate equations derived for the spherlcosymmetric model of a
burning drop. The weight fraction of fuel vapor at the surface of the
drop is given by the following equation (ref. 15):
Yf = 1
]l + — ?■ 1 / n J. ^~<^ J. \ 3- ~ §J_
m * s^
(5)
12
In using this equation for the computation of Tg, an iterative pro-
cedure was employed. First, the assumption was made that the tempera-
ture Tg was the normal boiling point of the fuel, and the weight
fraction of fuel vapor at the interface Y^ was calculated. For this
value of Yf the equilibrium temperature Tg was obtained from vapor
pressure data and substituted again into equation (5) as the next
approximation in the iteration. Rapid convergence of the results was
noted; for most of the fuels two repetitions produced a constant value
of Yf . As in the calculation of the theoretical burning-rate coeffi-
^ W
cients, mean values of the transport parameters were used. -■
2
The results are presented in table II together with the experi- g
mental values obtained. In general, there is good agreement.
Stanford Reseaxch Institute,
Menlo Park, Calif., January 15, 1958.
13
APPENDH
DETERMINATION OF DISTILLATION CURVES FOR
BINARY FUEL MIXTURES
The compositions of liquid and vapor in thermodynamic equilibrium
at the normal boiling points of the binary fuel mixtures employed in
W this investigation were determined with the apparatus of Daniels et al.,
1 reference Ik (fig. 12) . A distilland of known composition was placed
2 in the flask by dispensing known volimes of the constituents from
9. burettes . The fluid was brought to a boil and allowed to ref liix until
a steady boiling point was established. The apparatus was then allowed
to cool, and the reflux residue, which represented at most 1 percent of
the total liquid volume of the system, was removed with a pipette and
analyzed .
The temperatijre of the boiling distilland was measured with a pre-
cision of ±0,1° C with aja ASTM aniline point thermometer ( 51 -millimeter
immersion; range, 90°-170° C) . The pressure of the atmosphere varied
between 75^ and 76O millimeters of mercury during the coiirse of the
experiments .
Refliix residues from the heptane -butanol and the 2,2,i<—
trimethylpentane-butanol systems were analyzed by determining the
refractive indices of samples with an Abb^ refractometer . Analyses of
the dibutyl ether-pentanol system were made by means of vapor -phase
chromatography. Samples of the condensed vapor were introduced into a
column containing a silicone oil adsorbent maintained at 150° C. The
constituents were separated cleanly in the column, and their relative
proportions in the efflvix were determined with a thermal conductivity
cell which had been previously calibrated with known mixtures of the
compounds. The resulting distillation curves are presented in figure 15.
Ik
REFERENCES
1. Wise, H., and Agoston, G. A.: The Burning of a Liquid Droplet.
Literature of the Combustion of Petroleum, no, 20, Advances in
Chemistry Series, Amer. Chem. Soc, 1958> PP- 116-135 •
2. Blinov, V. J.: Change in Composition of Binary Liquid Mixtures on
Burning. Doklady Akad. SSSR, vol. 89, no. 1, 1953, PP. 101-10J4-.
3. Burgoyne, J. H., and Richardson, J. F.: Extinguishing Burning at a
Wick by the Use of Non-inflammable Liquids. Fuel, vol. 29, no. k,
Apr. 1950, pp. 95-97-
k. Godsave, G. A. E.: Studies of the Combustion of Drops in a Fuel
Spray - The Burning of Single Drops of Fuel. Fourth Symposium
(int.) on Combustion, Williams and Wilklns Co. (Baltimore), 1953,
pp. 818-830.
5. Wise, H., Lorell, J., and Wood, B. J.: The Effects of Chemical and
Physical Parameters on the Burning Rate of a Liquid Droplet. Fifth
Symposium (int.) on Combustion, Reinhold Pub. Corp., 1955,
pp. 132-114-1.
6. Agoston, George A., Wood, Bernard J., and Wise, Henry: Influence
of Pressure on the Combustion of Liquid Spheres. Jet Propulsion,
vol. 28, no. 5, Mar. 1958, pp. I8I-I88.
7. Wood, B. J., and Wise, H.: Measurement of the Burning Constant of a
Fuel Drop. Jo\ir. Appl. Phys.,. vol. 28, no. 9, Sept. 1957, P- IO68.
8. El Wakil, M. M., Priem, R. J., Brikowskl, H. J., I^ers, P. S., and
Uyehara, 0. A.: Experimental and Calculated Temperature and Mass
Histories of Vaporizing Fuel Drops. NACA TN 3^90, I956.
9. Smith, S. R., and Gordon, A. S.: Studies of Diffusion of Flames, I
and II. Jour. Phys. Chem., vol. 60, no. 6, June I956, pp. 759-763;
no. 8, Aug. 1956, pp. 1059-1062.
10. Perry, John H., ed.: Chemical Engineers' Handbook. McGraw-Hill Book
Co., Inc., 1950, p. 580.
11. Hottel, H. C, Williams, J. C., and Sinrpson, H. C: Combustion of
Droplets of Heavy Liquid Fuels, Fifth Symposium (int.) on Com-
bustion, Reinhold Pub. Corp., I955, pp. 101-129 ,
12. Berlad, A, H., and Hibbard, R. R.: Effect of Radiant Energy on Vapor-
ization and Combustion of Liquid Fuels, NACA RM E52IO9, 1952.
15
15. Lorell, J., Wise, H., and Carr, R. E.: Steady-State Biornlng of a
Liquid Droplet. II. Jour. Chem. Phys . , vol. 25, no. 2, Aug.
1956, pp. 325-351.
Ih . Daniels, F., Mathews, J. H., and Williajns, J. W.: Experimental
Physical Chemistry. McGraw-Hill Book Co., Inc., 19^4-1, pp. 79-81 <
16
TABLE I
BURNING CHARACTERISTICS OF DROPS OF MIXED FUELS^
Mixtures
Boiling point of
constituents,
OC
Character of
combustion
process
Increment between
boiling points
of terminal
constituents,
°C
Heptane
Nonane
98.4
151
Quiet
52.6
Heptane
Tetradecane
98.4
253
Disruptive
154.6
Heptane
Hexadecane
98.4
288
Disruptive
189.6
Heptane
Nonane
Tetradecane
98.4
151
255
Disruptive
154.6
Heptane
Octane
Nonane
Tetradecane
98.4
126
151
255
Disruptive
154.6
Nonane
Tetradecane
151
255
Quiet
102
Nonane
Hexadecane
151
288
Disruptive
137
Nonane
Tetradecane
Hexadecane
151
255
288
Quiet
137
Heptane
Nonane
Tetradecane
Hexadecane
98.4
151
255
288
Disruptive
189.6
Heptane
Octane
Nonane
Tetradecane
Hexadecane
98.4
126
151
253
288
Disruptive
189.6
w
1
2
9
liquid solution was made up of equal volumes of each
constituent .
17
TABLE II
SURFACE TEMPERATURES OF BURNING LIQUID DROPS
Fuel
Boiling
point.
Surface temperature
, °C
Theoretical,
(ref= 15)
Porous sphere
experiment
Hanging drop
experiment
Ethanol
78A
68
72.5
72
Butanol-1
118
101
108
105
Pentanol-1
138.1
117
124
125
Octanol-1
19^^.5
161
165
171
2,2' -Oxydiethanol
2kk
211
215
226
Heptane
98 A
81+
81+
82
Octane
125.7
—
108
1014-
Kerosene
^225
^185
202
^Mean boiling point.
Calculated on basis of a pure hydrocarbon with boiling point
of 2250 C.
18
TABLE III
BROPERTIES OF FUELS USED IN CALCUIATION OF
BURNING-RATE COEFFICIENT
[Transport properties of vapor fiLm were taken as those of air
at logarithmic mean temperature between flame and ambient
temperatiore (cf. pp. 1 and 6, ref. 6): Cg = O.27I cal/gm °C;
Ag = 1.56 X 10"^ cal/sec cm °C. Since drops were "biirned in
air at 1 atmosphere, Yg^ = 0.232 J
Fuel
gm/cmP
cal/gm °C
cal/gm
gm/gm
q/i,
cal/gm
hp,
°C
Butanol-1
0.810
207.0
8,051
2.59
3,110
118
Heptane
0.681+
121.6
10,720
5.51
3,060
98
2, 2,i^-Trimethylpentane
0.692
106.8
10,672
5.51
5,01+0
99
Pentanol-1
0.818
201
8,1+20
2.72
5,100
158
Dibutyl ether
0.773
185
8,850
2.87
5,080
1I+2
19
TABLE IV
COMPOSITION VARIATION OF BINARY FUEL DROPS DURING COMBUSTION
SYSTEM: 2,2,4-TRIMETHYLPENTANE-BUTANOL-l
Elapsed time after
Composition, mol percent '
Initial drop diameter.
ignition, sec
Butanol-1
cm
10
• 5
8.5
1.0
6.0
1.5
3.5
l6.2 X 10-^
2.0
2.0
2.5
2.6
1^0
.5
k6
1.0
58
1.5
70
iD.l X 10-2
2.0
100
■
2.5
100
2.75
100
50
.5
55
1.0
65
1.5
88
15.6 X 10-2
2.0
100
2.5
100
3.0
100
5.2
100
60
.5
66
1.0
85
15.2 X 10-2
1.5
100
2.0
100
2.5
100
3.0
100
75
.5
81
1.0
91
1-5
100
2.0
100
16.1 X 10-2
2.5
100
3.0
100
3.5
100
3.75
100
20
a, b PURE FUELS
C BINARY FUEL MIXTURE
TIME
(a) Nonequilibriurn distillation.
a
I
I-
1 1 1 i 1 1 1
e, f PURE FUELS
<
^ g BINARY FUEL MIXTURE —
cr
<
1
>^\
<
z
o
^^
1-
o
UJ
W)
1
en
en
O
a:
o
^^X \
, , , ,\ ^ ^
TIME
(b) Equilibrium distillation.
Figure 1.- Schematic presentation of variation of drop cross-sectional
area with time during combustion.
21
ON
OJ
to
:i
-p
a
u
cs
p<
ft
ft
I'
o
05
to
-P
o
0)
H
(1)
O
-P
o
ft
o
0)
•H
o
•H
(U
OJ
OJ
(in
22
FLAME
OUTLINE
TC TO REFERENCE
JUNCTION AND RECORDER
TO SERVO-CONTROLLED
FUEL FEED DEVICE
POROUS SPHERE
(ALUNDUM)
THERMOCOUPLE
JUNCTION
Figxire 5.- Schematic diagram of fuel-wetted porous sphere with thermo-
couple for measurement of liquid surface temperatures.
23
TIME, 0.5-SECOND INTERVALS
Figure k.- Time -temperature record associated with burning of suspended
enthanol fuel drop.
2k
4eopxio"^ p
TIME, l-SECOND INTERVALS
(a) Butanol-1 and heptane,
TIME, I -SECOND INTERVALS
(b) Butanol-1 and 2,2,4-trimethylpentane.
rxio" 1 1 1 1 1 1 I
INITIAL COMPOSITION - MOL PERCENT PENTANOL-
1 r
TIME, l-SECOND INTERVALS
(c) Pentanol-1 and dibutyl ether.
Figure 5.- Burning-rate curves for various binary fuel mixtures,
25
.01 I
.012
.0 1 I
.010 -
.009 -
.008 -
.007
INITIAL
FINAL
• NONEOUILIBRIUM DISTILLATION •
(THEORETICAL)
25 40 50 60 75
COMPOSITION, MOL PERCENT BUTANOL-I
(a) Heptane -butanol-1.
100
INITIAL
FINAL
■ NCNEaUiLIBRMJM DISTILLATION
(THEORETICAL)
10 17.5 25 40 50 60 75
COMPOSITION, MOL PERCENT BUTANOL-I
(b) 2,2,4-trimethylpentane-butanol-l .
90 100
.Oil
.010
.009
.008
INITIAL
FINAL
• NONEOUILIBRIUM DISTILLATION
•i (THEORETICAL)
25 40 50 61.5 67.5 75
COMPOSITION, MOL PERCENT PENTANOL-I
85
100
Figure 6.-
(c) Dibutyl ether-pentanol-1.
Burning-rate coefficients for various binary fuel mixtures,
26
<
UJ
UJ
en
I
(/)
O
q:
o
■J^^^^^^tfuft^A^ 4 nMynMrmihuiniOiiiiiMiiiuifl n dtttR^^^^^^nrasHSj dS^^^^a^H^KSft^ ^^^^ffiH^ttS^K ^^GtfiHttlBlffiBI^^ ||^^||H|uj||^Ha^te ^Ag^f^g^^^^fluyj. d|Uuwmj^^tiygtf|i^ jTtfgMMjrtMMHmffljilg^'
l^iMBW w^iiK » » ^^ ^ -^ W/tt 'WKM DISRUPTION SHOWN B
■MB HbiS"!'^
TIME, 0.5-SECOND INTERVALS
Fig\ire 7-- Burning-rate curve for binary fuel showing disruptive combus-
tion. (Fuel: 50-50 mol percent heptane -hexade cane ^ Gf . fig, 11.)
I
ro
27
On
OJ
tH
I
—
1
VAPOR
1
-^
111
—
/-
z>
^^^^
/^
h-
^^^^
<
^^^
^^^
ac
^^^
A ^*»^
UJ
Q.
>^^ -C
UJ
1-
~ y
B^^^^^
—
^
^^
LIQUID
1
—
COMPOSITION
Figure 8.- Schematic presentation of vaporization of binary fuel drop
during combustion. Nonequilibrium distillation. A, C; equilibrium
distillation, B.
28
TIME, I -SECOND INTERVALS
(a) Butanol-1 and heptane.
\ \ \ \ \ \ r
INITIAL COMPOSITION - MOL PERCENT BUTANOL-I
TIME, I - SECOND INTERVALS
(b) Butanol-1 and 2^2^4-trimethylpentane.
480X10"* -
T \ 1 \ \ \ I
INITIAL COMPOSITION - MOL PERCENT PENTANOL-
1 r
J L
J L
TIME. I - SECOND INTERVALS
(c) Pent ano 1-1 sind Dibutyl ether.
Figure 9.- Theoretical b\irning-rate ciirves for various mixtures.
Circles indicate experimental points.
29
ON
OJ
o
CVI
ro
o
o
o
00
't
o
CVJ
CJ
CVJ
o
CO
o
CVJ
O
00
-=1-
0)
:i
+3
0)
t:)
03
0)
Q)
CO
O
U
(U
AJ
fH
O
to
0}
§>
•H
I
giijo *a3avnos aaiHwvia doaa
50
t = l.l49
seconds
t= 1.182
seconds
in
t = 1.199
seconds
t = 1 . 233
seconds
m
t= 1.250
seconds
3E
t = 1.267
seconds
IZ
f =1. 216
seconds
T t
t = 1.264
seconds
L-59-3098
Figure 11.- Formation of bubble in liquid phase of burning drop. (Fuel:
5O-5O mol percent heptane -he xade cane; diameter of supporting fiber
tip: 1,000 microns; t indicates elapsed time after ignition;
burning-rate curve produced by this drop is presented in fig. 70
51
A.S.T.M.
ANILINE-POINT-
THERMOMETER
TO VARIAC
ON
OJ
'"OTOOOTTS^r^i
TOWOOWO';
Co CI2 TUBE
^TO DRAIN
REFLUX RESIDUE
NICHROME HEATER
Figure 12.- Apparatus for determination of boiling points and liquid -
vapor compositions of binary solutions .
52
(a) EeptEme-butanol-l.
J [ L
(b) 2,2,4-trimethylpentane-butanol-l.
(c) Dibutyl ether-pentanol-1.
Figure I5.- Liquid-vapor equilibrium composition as a function of
temperature at a pressure of 1 atmosphere .
NASA - Langley Field, Va. W-129